Questions and Answers :
GPU applications :
Rather newbie question
Message board moderation
Author | Message |
---|---|
Jens Send message Joined: 8 Jun 99 Posts: 51 Credit: 153,539 RAC: 0 |
Sorry for this rather stupid question, but now that CUDA is working, I have the following situation: After creating a cc_config-file (I have a dual-core processor), there are two tasks calculated by the CPUs (AP) and one by the GPU (S@H). I suppose, that's the way it's meant to be? The problem is, that the two AP-units seem to be calculated at a very low speed. Now they are estimated to be finished in 180 hrs. If I'm not mistaken normally it was 80. Am I doing something wrong? The problem: When I do it without the cc_config-file, then only 50% of my CPU is used. Best regards Jens |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
Most likely you did all just right. It's BOINC error in estimation of time to complete. Try to look at completion time and % done after few hours. AP has pretty linear processing advance (I mean % per hour pretty stable) so you could estimate completion time by yourself. Hope it will be much less than BOINC thinks (it improve its predictions too after few AP task completion). BTW, do you use opt AP or stock one? |
Gundolf Jahn Send message Joined: 19 Sep 00 Posts: 3184 Credit: 446,358 RAC: 0 |
...The problem: When I do it without the cc_config-file, then only 50% of my CPU is used. Nah, it's about 50.3% to 50.6% ;-) (And with your 2+1 configuration, those 3-6% might be missing for feeding the GPU :-) Gruß, Gundolf |
Jens Send message Joined: 8 Jun 99 Posts: 51 Credit: 153,539 RAC: 0 |
Most likely you did all just right. It's BOINC error in estimation of time to complete. Hey Raistmer, well - looks right know, like I'd need about 120 hrs per AP-WU. This is significantly more than before, if I'm not mistaken. Any idea on what might be the reason? I use the AP, that was sent to me by SETI - Astropulse 5.00 Best regards Jens |
Geek@Play Send message Joined: 31 Jul 01 Posts: 2467 Credit: 86,146,931 RAC: 0 |
I'm seeing the same thing. 20% to 25% increase in the AstroPulse actual crunch times than before CUDA. Boinc....Boinc....Boinc....Boinc.... |
Geek@Play Send message Joined: 31 Jul 01 Posts: 2467 Credit: 86,146,931 RAC: 0 |
Uhm.......... This morning the AstroPulse times seem normal. Perhaps my eyes got crossed up last night. Boinc....Boinc....Boinc....Boinc.... |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.